[whatmodel.sh](../../README.md) / Phi-3.5-MoE-instruct
# Phi-3.5-MoE-instruct
Microsoft · released 2024-08 · 41.87B params (6.6B active) · mit
**score 49.2** ±1.1 · #387 of 732 · from 6 benchmark results across 4 sources
Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only.
## What it's good at
● marks Phi-3.5-MoE-instruct among every model on that task or test, worst to best.
other tests (6)
test result vs the others in its run difficulty counts why
LiveBench 2024-07-26 35.2% ▁▁▁▂▂▃▂█●▃▅▃ beats 38/61 medium · avg 38% 50% at its level
CLIcK Korean (Daekeun Kim) 56.4% ▆●▄▄█▄▂·▄█·▄ beats 4/23 medium · avg 56% 21% at its level
HAE-RAE 1.0 (Daekeun Kim) 61.8% ▄▄●▂▂█▂▄▂█▂▄ beats 5/22 medium · avg 61% 18% at its level
BRIDGE clinical text, zero-shot a… 29.5% ▁▁▁▁▁▂▂●▄▆█▃ beats 30/97 hard · avg 31% 6.8% blunt · at its level
KMMLU-HARD (Daekeun Kim, zero-sho… 25.3% ●···▁▁▁▂▃·▆▂ beats 3/23 very hard · avg 29% 3.1% sharp · too hard for it
Vectara hallucination leaderboard… 97.5% ▁▁▂▃▄▄█▆●▂▂▁ beats 115/140 very easy · avg 94% 0.8% blunt · maxed out · noisy
source results share of the score
LiveBench model judgments 1 50% ▇▇▇▇▇▇
Daekeun Kim Korean LLM benchmark runs 3 42% ▇▇▇▇▇
BRIDGE Medical Leaderboard 1 7% ▇
Vectara hallucination leaderboard 1 1% ▇
prior: pull toward the middle 1% ▇
weighted average = 49.2
Licences: LiveBench model judgments ([Apache-2.0 (stated in the LiveBench datasheet)](https://www.apache.org/licenses/LICENSE-2.0)) · Daekeun Kim Korean LLM benchmark runs ([MIT](../../licenses/korean-personal-benchmark-runs-MIT.txt)) · BRIDGE Medical Leaderboard ([MIT](../../licenses/bridge-medical-leaderboard-MIT.txt)) · Vectara hallucination leaderboard ([Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0))
## How to run it
### llama.cpp (recommended)
1. Install llama.cpp. Homebrew and winget installs name the commands llama-cli and llama-server instead. [Other ways to install](https://github.com/ggml-org/llama.cpp/blob/master/docs/install.md)
```sh
# macOS / Linux
curl -LsSf https://llama.app/install.sh | sh
# Windows
irm https://llama.app/install.ps1 | iex
```
2. Run Phi-3.5-MoE-instruct (Q4_K_M, needs ~25.3 GB, from [huggingface.co/lmstudio-community](https://huggingface.co/lmstudio-community/Phi-3.5-MoE-instruct-GGUF/tree/ae1ea4c56c6bd7fc73ab92778a11e70365a314e2))
```sh
llama cli -hf lmstudio-community/Phi-3.5-MoE-instruct-GGUF:Q4_K_M
```
3. Or start a local server and open the address it prints (http://localhost:9931)
```sh
llama serve -hf lmstudio-community/Phi-3.5-MoE-instruct-GGUF:Q4_K_M
```
Other versions: Q8_0 (~44.5 GB) `llama cli -hf lmstudio-community/Phi-3.5-MoE-instruct-GGUF:Q8_0`
### LM Studio
1. Install LM Studio. Open the app once before using the lms command. [Download it (macOS / Windows / Linux)](https://lmstudio.ai/download)
2. Run Phi-3.5-MoE-instruct (Q4_K_M, needs ~25.3 GB, from [huggingface.co/lmstudio-community](https://huggingface.co/lmstudio-community/Phi-3.5-MoE-instruct-GGUF/tree/ae1ea4c56c6bd7fc73ab92778a11e70365a314e2))
```sh
lms get https://huggingface.co/lmstudio-community/Phi-3.5-MoE-instruct-GGUF@Q4_K_M -y
```
3. Then run lms ls to see its name, and lms chat followed by that name
Other versions: Q8_0 (~44.5 GB) `lms get https://huggingface.co/lmstudio-community/Phi-3.5-MoE-instruct-GGUF@Q8_0 -y` · 4bit (~23.6 GB, Mac only) `lms get https://huggingface.co/mlx-community/Phi-3.5-MoE-instruct-4bit -y` · 8bit (~44.5 GB, Mac only) `lms get https://huggingface.co/mlx-community/Phi-3.5-MoE-instruct-8bit -y`
### Ollama
1. Install Ollama. [Download it (macOS / Windows)](https://ollama.com/download)
```sh
# Linux
curl -fsSL https://ollama.com/install.sh | sh
```
2. Run Phi-3.5-MoE-instruct (Q4_K_M, needs ~25.3 GB, from [huggingface.co/lmstudio-community](https://huggingface.co/lmstudio-community/Phi-3.5-MoE-instruct-GGUF/tree/ae1ea4c56c6bd7fc73ab92778a11e70365a314e2))
```sh
ollama run hf.co/lmstudio-community/Phi-3.5-MoE-instruct-GGUF:Q4_K_M
```
Other versions: Q8_0 (~44.5 GB) `ollama run hf.co/lmstudio-community/Phi-3.5-MoE-instruct-GGUF:Q8_0`
### MLX (Mac)
1. Install MLX (Mac). Apple silicon Macs only, text only.
```sh
# with uv
uv tool install mlx-lm
# or
pip install mlx-lm
```
2. Run Phi-3.5-MoE-instruct (4bit, needs ~23.6 GB, Mac only, from [huggingface.co/mlx-community](https://huggingface.co/mlx-community/Phi-3.5-MoE-instruct-4bit/tree/c55613aaad97402ea5500941a0ea458e63714a57))
```sh
mlx_lm.chat --model mlx-community/Phi-3.5-MoE-instruct-4bit
```
3. Or start an OpenAI-compatible server at http://localhost:8080/v1
```sh
mlx_lm.server --model mlx-community/Phi-3.5-MoE-instruct-4bit
```
Other versions: 8bit (~44.5 GB, Mac only) `mlx_lm.chat --model mlx-community/Phi-3.5-MoE-instruct-8bit`
### Other ways to run it
1. Run it as published (BF16, needs ~100.99 GB, NVIDIA only, from [huggingface.co/microsoft](https://huggingface.co/microsoft/Phi-3.5-MoE-instruct))
```sh
vllm serve microsoft/Phi-3.5-MoE-instruct
```
Model repository and documentation: [huggingface.co/microsoft](https://huggingface.co/microsoft/Phi-3.5-MoE-instruct)
[interactive version](https://whatmodel.sh/model/phi-3-5-moe-instruct) · [JSON](https://whatmodel.sh/api?format=json&all=1)